Finite Element Simulation of Two Vowel Changes in a Male Vocal Tract Model With an Aramany Class I Maxillectomy

Acoustic analysis of vowels offers an objective method for evaluating speech in maxillectomy patients. Prior studies have focused primarily on lower formants linked to vowel identification, while changes in higher formants-critical for voice quality and individualization-remain underexplored.

This study used finite element (FE) modeling to analyze vocal tract transfer function up to 10 kHz, acoustic mode shapes, and phonation of the vowels /aː/ and /iː/ in Aramany class I maxillectomy models.

Healthy and maxillectomy vocal tract models were created using magnetic resonance imaging data from the Dresden Vocal Tract Dataset (DVTD). Vocal tract transfer function analysis was performed using FE modeling in ANSYS Workbench, stimulating the vocal tract at the glottis and recording the frequency spectrum at the mouth. Acoustic mode shapes corresponding to spectral changes were extracted. Time-domain phonation was simulated using Liljencrants-Fant glottal pulses. The results were validated against experimental data from the DVTD.

Maxillectomy models showed notable spectral alterations, including shifts in both lower and higher formants as well as the emergence of new peaks and dips. Changes in the first three formants aligned with previous clinical findings. Mode shape analysis revealed that defect regions contributed to specific transfer function peaks changes. Phonation simulations were successful in replicating vowel sounds.

In addition to lower formants, maxillectomy defects disrupt higher resonance frequencies linked to speech individuality. This highlights the complex nature of maxillectomy speech and underscores the need for detailed acoustic tools to guide personalized prosthetic and speech rehabilitation. Further research is needed to explore other defect classes and their perceptual impacts.